In the hippocampal formation in vivo, brief periods of gamma-frequency activity follow population bursts called sharp waves. The approximately 200 Hz activity of the sharp wave itself may serve to enhance synaptic connections and the approximately 40 Hz gamma activity has been offered as a mechanism for solving the 'binding' problem. We describe epochs of gamma-frequency activity which follow population spikes evoked by low frequency repetitive extracellular stimuli in retrohippocampal neurons of horizontal rat brain slices. 2. gamma-Frequency activity recorded intracellularly from deep layer neurons of entorhinal cortex, presubiculum and parasubiculum consisted of one action potential correlated with each of the three to five gamma cycles recorded with a proximate field potential electrode. A minority of cells exhibited only sub-threshold gamma-frequency membrane potential oscillations (ranging from 5 to 10 mV). No cells fired more than one spike per gamma cycle under any conditions. 3. The range of synchrony varied from individual cells which showed gamma-frequency firing without corresponding oscillations in close field recordings to field potential recordings of oscillations which were well correlated across regions. The lead or lag between any two retrohippocampal regions was in the direction of the conduction delay for the primary population spike, but typically was less, and approached zero milliseconds for some cycles in most cells. The level of synchrony was stable for particular stimulating conditions (intensity, stimulation rate, stimulus location). 4. The duration of the period of gamma activity had the duration of a slow depolarizing potential which was mediated by NMDA receptor activation. NMDA receptor antagonists or low concentrations of AMPA receptor antagonists reduced the duration of, or completely abolished the slow potential, thereby eliminating the gamma portion of the evoked response. 5. gamma-Frequency firing was eliminated by the GABAA receptor antagonist picrotoxin but small (< 5 mV) membrane potential oscillations remained after focal picrotoxin applications, and these exhibited the voltage dependence of EPSPs. Bath application of thiopental lowered the frequency of gamma oscillations, confirming the involvement of GABAA receptors. 6. The GABAB receptor antagonist 2-hydroxy-saclofen appeared to enhance the gamma activity by increasing the duration of the gamma epoch and increasing the amplitude of individual gamma cycles in field potential recordings. These saclofen-induced cycles were, however, less well synchronized across regions. 7. We show that synchronous gamma (40-100 Hz) activity follows population bursts by deep layer retrohippocampal neurons in undrugged slices from rat brain. Responses from medial entorhinal, parasubicular or presubicular cells were not distinguishable. These events can be initiated by a propagating population spike. We suggest that an NMDA receptor mediated depolarization enables the network of deep layer retrohippocampal neurons to oscillate by providing a sustained excitation, the duration of which determines the duration of the gamma episode. gamma-Frequency firing is primarily the result of GABAA receptor dependent inhibition during this period of sustained depolarization. Recurrent excitation appears to be inconsequential for principal cell firing, but may contribute to interneuron firing.
